基于UKF方法的永磁同步平面电动机位移测量

Permanent-magnet Synchronous Planar Motor Displacement Measurement Based on UKF Method

针对现有永磁同步平面电动机位移测量方法测量自由度少,且安装要求苛刻等缺点,利用少量分散排布,在磁场中相位不重合的霍尔传感器组成的霍尔阵列信号及电动机的磁通密度模型,基于无味卡尔曼滤波(Unscented Kalman filter,UKF)方法导出一种永磁同步平面电动机沿x、y轴平动及绕z轴转动位移的测量方法。仿真及试验结果表明,该方法可以实现电动机平面3自由度位移的测量。在观测值使用次数一定的情况下,随着观测噪声逐渐减小,位移估计值精度便逐渐提高,但当观测噪声小到一定程度时,位移估计值精度稳定在一定水平上,此时由算法造成的计算误差成为影响精度的主要因素。在同样的观测噪声下,多次使用观测值进行反复迭代,可以使位移估计值精度得到进一步提高。在实际工程应用中,应该从精度要求和计算量综合考虑,从而确定最优观测值迭代次数。

Existing permanent-magnet synchronous planar motor’s displacement measurement methods have shortcomings such as insufficient degrees of freedom and demanding installation.Based on the unscented Kalman filter（UKF） method,a new measuring method is derived by using a motor magnetic flux density model and the signal from a Hall-array composed of a few dispersedly placed Hall sensors which have different phases in magnetic field.Simulation and experiment results show that the above method can measure permanent-magnet synchronous planar motor’s translational displacement along x and y axes,and rotational displacement around z axis.Moreover,at the same observation iterations,with the gradual decrease of observation noise,the accuracy of displacement estimates becomes higher and higher,but when the observation noise decreases to a certain extent,the accuracy stabilizes at a certain level.Then the calculation error caused by the algorithm becomes the main factor affecting the accuracy.At the same observation noise,using observation values repeatedly can improve the accuracy of displacement estimates.However,in practical engineering applications,the required accuracy and calculation amount should be considered comprehensively,then the optimal observation iterations can be determined.